The key idea when using switched antenna diversity is to have more than one antenna in the receiver. For each received frame the signal level of the received signal, in particular the received signal strength (RSS) is measured for all or a selection of the antennas in the receive path. These RSS measurements are made during the initial preamble part of the frame that is used for receiver training. Before the data part of the frame is entering the receiver, the antenna with the strongest RSS is selected for reception of the rest of the frame. In this way an increased receiver performance will be gained compared to using only one antenna.
During the preamble different tasks have to be performed in the receiver. The most important task is to adjust the signal level to be within the dynamic range of the receiver by setting the gain of the amplifier. This is done by measuring the RSS of the amplified signal and adjusting the gain of the amplifier repeatedly until the signal level is in accordance with the dynamic range of the receiver. This procedure is done using one (i) of the N antennas of the station.
After the gain of the amplifier has been set, the antenna diversity measurements are performed. The last RSS measurement done during the gain setting can be used as RSS estimate for antenna i. For the remaining antennas separate RSS measurements must be performed.
For simplicity it can be assumed that the station comprises two antennas where antenna 1 is used for the gain calculations. There are two situations which can arise after the RSS measurements have been made for both antennas:                1. RSS for antenna 1 is higher than for antenna 2 (RSS1>RSS2)        2. RSS for antenna 1 is higher than for antenna 1 (RSS2>RSS1)        
In case 1 above the gain setting is already as good as it can be since the best antenna is the same antenna as that has been used for gain calculations. In case 2 the antenna not used for gain calculations is found to be the best one and if the difference between the two antennas is large we will need to recalculate the gain again otherwise we might experience a reduced performance compared to always use antenna 1.
The problem is that gain recalculations take time and time is a very limited resource during the preamble.
A further problem of antenna diversity lies in the following. When using more than one antenna in the receiver it is possible that the noise level on all antennas is not equal. The cause of this can be a bad placement of the antennas or that one antenna experiences stronger adjacent channel disturbance. With weak signal levels the received noise level will be so high that it heavily affects the antenna selection. We will choose the antenna with the highest noise level instead of the one with the strongest signal level which would be the correct choice.
The best solution would be if it was possible to estimate the noise level for each received frame and take this into account when selecting antenna. This is however both difficult and expensive in terms of hardware costs.